Bioinformatics analysis and verification of gene targets for benign tracheal stenosis.

BACKGROUND: Tracheal injury could cause intratracheal scar hyperplasia which in turn causes benign tracheal stenosis (TS). With the increasing use of mechanical ventilation and ventilator, the incidence of TS is increasing. However, the molecular mechanisms of TS have not been elucidated. It is significant to further explore the molecular mechanisms of TS.
METHODS: The repeatability of public data was verified. Differently expressed genes (DEGs) and most significant genes were identified between TS and normal samples. Enrichment analysis of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) were analyzed. The comparative toxicogenomics database were analyzed. TS patients were recruited and RT-qPCR were performed to verify the most significant genes.
RESULTS: There exist strong correlations among samples of TS and normal group. There was a total of 194 DEGs, including 61 downregulated DEGs and 133 upregulated DEGs. GO were significantly enriched in mitotic nuclear division, cell cycle, and cell division. Analysis of KEGG indicated that the top pathways were cell cycle, and p53 pathway. MKI67(OMIM:176741), CCNB1(OMIM:123836), and CCNB2(OMIM:602755) were identified as the most significant genes of TS, and validated by the clinical samples.
CONCLUSION: Bioinformatics methods might be useful method to explore the mechanisms of TS. In addition, MKI67, CCNB1, and CCNB2 might be the most significant genes of TS.

INTRODUCTION

Tracheal stenosis (TS) refers to the reduction in diameter of the trachea, caused by tracheotomy, intratracheal tumors, tracheomalacia, infections, inflammations, and other pathologic conditions (Farzanegan et al., 2017; Raghuraman, Rajan, Marzouk, Mullhi, & Smith, 2005). What is more, most cases of TS develop as a result of tracheal intubation or tracheostomy which induce intratracheal scar hyperplasia (Kim, Khalpey, Hsu, & Little, 2017). A retrospective study of 150 patients with TS showed that 54.7% of the cases are caused by iatrogenic injury, 18.5% are caused by autoimmune, and 8% are caused by trauma (Gelbard et al., 2015). In addition, with the increasing use of mechanical ventilation and ventilator, the incidence of TS is increasing (Ahn, Su, Kim, & I H., 2015). Due to dyspnea, patients’ quality of life is badly affected. When symptoms are severe, surgical intervention is warranted (Farzanegan et al., 2017).

Multiple mechanisms are involved in tracheal stenosis, but the molecular mechanisms remain unclear. The endotracheal tube cuff excessively compresses the tracheal wall, causing mucosal injury and inflammatory response, which in turn induces tracheal wall damage and scar formation (Su et al., 2017). Meanwhile, inflammatory cells and immune cells caused by tracheal wall damage secrete a variety of inflammatory factors and growth factors such as IL‐6 which can promote fibrous scar formation (Yin et al., 2017). Furthermore, resection of the stenotic lesions, balloon dilation, and stenting are the main treatments to relieve stenosis (D'Andrilli, Venuta, & Rendina, 2016). However, the surgical treatment itself might also cause new tracheal injury and anastomotic complications, inducing new fibrous scar (Wright et al., 2019). Generally, early diagnosis and management for the TS could bring a better prognosis to the patients (Smith & Cotton, 2018). Therefore, it is significant to further explore the mechanisms of TS and find molecular targets that can be used to diagnose early, prevent early, and treat early.

Bioinformatics methods are widely used to find molecular changes in the occurrence and development of diseases and are effective ways to explore the pathogenesis of diseases. Zhang found the genes related to the pathogenesis of oralsquarous cell carcinoma (OSCC) utilizing bioinformatics analysis, and further verified these molecules might lead to OSCC through the PI3K‐AKT pathway, suggesting the relevant molecules may be a molecular target for specific diagnosis and therapy (Zhang, Feng, et al., 2018; Zhang, Zhang, et al., 2018). In addition, Liu found abnormally expressed molecules in multiple scar tissues by bioinformatics analysis and further verified that SFRP1(OMIM:604156) might take part in the occurrence of scar by regulating Wnt/β‐catenin (Liu et al., 2018). However, there are few reports on the use of bioinformatics analysis to explore the mechanisms of acute tracheal injury and TS.

Plenty of DEGs were identified between TS and control individuals via the bioinformatics analysis. Enrichment analysis was implemented. Furthermore, the research constructed networks of Protein‐protein Interaction (PPI), significant modules, and hub genes. Finally, three most significant genes (MKI67(OMIM:176741), CCNB1(OMIM:123836), and CCNB2(OMIM:602755)) were identified and verified by the functional experiment of the clinical samples.

MATERIALS AND METHODS

Obtaining the public data

The GEO (http://www.ncbi.nlm.nih.gov/geo) is an open source platform for the storage of genetic data (Edgar, Domrachev, & Lash, 2002). One expression profiling dataset [GSE109365 (GPL16570 platform)] were downloaded. The GSE109365 dataset (Musah et al., 2019) includes five tracheal stenosis samples which were exposed to chlorine 4 days, and five normal trachea which were not exposed to chlorine.

The repeatability test for the intra‐group data

The repeatability of data was verified by the Pearson's correlation test. R (Lin et al., 2017) is an open language and environment for statistical computing and mapping, which was maintained by a large and active global research community. The Pearson's correlation test and the mapping of heat maps were completed by the R language. Principal component analysis (PCA; Ringner, 2008), a strong mathematical method, was capable of reducing the data's complexity. The PCA could capture variance in the whole fields by detecting the linear combinations, so that the components, which were orthogonal to and not correlated with each other, were divided. Also, the repeatability of data was verified by the PCA.

Differently expressed genes (DEGs) identified by LIMMA package

As a fully functional package, the LIMMA package includes the original data input and preprocessing capabilities of chips, as well as a liner model for analyzing differentially expressed genes. We screened DEGs between normal and TS samples by utilizing LIMMA package with a adjust p < .001 and a log (Fold Change) ≥ 1 or log (Fold Change) ≤ −1. And the volcano plot was drawn by R language. The DEGs are presented as volcano plots, generated using SangerBox software (http://sangerbox.com/).

Construction of a PPI network

Search Tool for the Retrieval of Interacting Genes (http://string.embl.de/; Szklarczyk et al., 2015), an open source online tool, was used to construct a PPI network, and Cytoscape visualization software version 3.6.1(Smoot, Ono, Ruscheinski, Wang, & Ideker, 2011) was implemented to complete visualization. A confidence score >0.4 was set, which may filter out the critical module.

GO and KEGG analysis via DAVID tool

One online tool, DAVID (Huang et al., 2007; https://david.ncifcrf.gov/home.jsp; version 6.8, Maryland, America), was applied to carried out the functional annotation for DEGs. GO (Ashburner et al., 2000) generally perform enrichment analysis of genomes. And there are mainly cellular components (CC), biological processes (BP), and molecular functions (MF) in the GO analysis. KEGG (https://www.kegg.jp/; Tanabe & Kanehisa, 2012) is a comprehensive database of genomic, chemical, and systemic functional information. Therefore, DAVID was used to make analysis of GO and KEGG.

Enrichment analysis by Metascape

Metascape (http://metascape.org/gp/index.html#/main/step1; Zhou et al., 2019) is a powerful annotation analysis tool for gene function, which can help researchers apply the current popular bioinformatics analysis methods to the analysis of batch genes and proteins, so as to realize the cognition of gene or protein functions. It can annotate a large number of genes or proteins. It integrates several authoritative functional databases, such as GO, KEGG, and Uniprot, to analyze not only human data, but also the data of many other species, and to analyze not only a single dataset, but also multiple gene sets simultaneously. The Matascape was performed.

The screening of significant module, and identification of hub genes

Molecular Complex Detection tool (MCODE; version 1.5.1; Bader & Hogue, 2003), one plug‐in of Cytoscape, could screen and identify the most significant module in the PPI network, and the criteria were given as the degree of cut‐off = 2, MCODE scores > 5, k‐score = 2, node score cut‐off = 0.2, and maximum depth = 100. Furthermore, once the degree was more than 10, the cytoHubba (Chin et al., 2014), one plug‐in of Cytoscape, could identify the hub genes.

The analysis of hub genes

The R was used to perform the clustering analysis based on the gene expression level. The Pearson's correlation test was performed to complete the correlation analysis among the hub genes. And the mapping of heat maps, which could present the correlation among the all hub genes, were completed by the R language.

The validation of most significant genes

The MCC algorithm of cytoHubba was also implemented to identify the hub genes. And the VENN was used to obtain the common genes between “cytoHubba_Degree” and “cytoHubba_MCC.” Then, the most significant genes were identified.

Identification of the most significant genes associated with trachea

The comparative toxicogenomics database (http://ctdbase.org/) is a web‐based tool that provides their relationships with diseases. The relationships between the significant genes and “Tracheal Diseases, Cartilage Diseases, Tracheal Neoplasms, Fibroadenoma, and Connective Tissue Diseases” were analyzed via comparative toxicogenomics database.

Ethical compliance

The research conformed to the Declaration of Helsinki and was authorized by the Human Ethics and Research Ethics Committees of the Second Hospital of Hebei Medical University. All participates provided the informed consents.

Patients

This study recruited a total of 25 individuals including 11 TS patients and 14 un‐TS individuals between 2017 and 2020 from the Second Hospital of Hebei Medical University. There was no difference in tube duration, and the balloon pressure was monitored during tube duration, and the pressure was within the clinically acceptable range. Clinical characteristics and basic information were available for all patients, and were collected retrospectively from medical records. Patients aged 18–100 years old, suspected tracheal stenosis TS based on symptom, not received treatment, and no history of tracheal surgery will be screened for inclusion criteria. Exclusion criteria included: age <18 years old or >100 years old, combined with other malignant diseases, and severs heart disease and diabetes.

RT‐qPCR assay

RNAiso Plus (Trizol) kit (Thermofisher) was implemented to extracte RNA of the most significant genes, including MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10). Light Cycler® 4800 System was used to make RT‐qPCR for the 10 hub genes. The primer sequences are presented in Table 1. The relative quantitative values (ΔCT) were calculated. The endogenous control was GAPDH.

TABLE 1

Primers and their sequences for PCR analysis

Primer	Sequence (5′–3′)
MKI67‐hF	GCTCACGCCTGTAATCCC
MKI67‐hR	TGCTCTTCGCTTTGCTTT
CCNB1‐hF	GTGTTCTAACTTTGGAGGAT
CCNB1‐hR	TCAGTCACCCAATACCAG
CCNB2‐hF	TAATGGGCATTTCTGTAAG
CCNB2‐hR	CAGGGCTATTGTTTGGTA

MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10).

Statistical analysis

The data were expressed as percentage of total and mean ± SD. Student's t test was used to compare value between two groups.

By using the Pearson's correlation test, associations between the degree of TS and the expression of MKI67, CCNB1, and CCNB2 were analyzed. And we used the linear regression analysis to explore the linear correlations between them. The Spearman‐rho test was used to compare degree of TS, the expression of MKI67, CCNB1, and CCNB2 for the correlation analysis.

RESULTS

High repeatability of data in the GSE109365

There existed strong relevance among individuals in TS group, and in the normal individuals in the GSE109365 via the Pearson's correlation test (Figure 1a). Furthermore, through the PCA, the repeatability of the data in GSE109365 was fine (Figure 1b).

FIGURE 1

(a) Pearson's correlation test manifests that there exist powerful correlations among samples in the TS group, and that there also exist powerful correlations among samples in control group in the GSE109365 dataset. (b) After performing the principal component analysis, the repeatability of data in GSE109365 was fine. (c) Volcano plot presents the DEGs between TS and control groups. (d) PPI network of the DEGs

DEGs between TS and normal samples and the PPI network

There was a total of 194 DEGs between TS and normal samples, including 61 downregulated DEGs and 133 upregulated DEGs. The volcano plot presents the DEGs in the GSE109365 (Figure 1c). And the PPI network was constructed (Figure 1d).

Functional annotation for DEGs via DAVID

The BP were mainly enriched in mitotic nuclear division, cell cycle, cell division, chromosome segregation, mitotic cytokinesis, and positive regulation of fibroblast proliferation (Figure 2a). The CC were mainly enriched in chromosome, centromeric region, and condensed chromosome kinetochore (Figure 2b). The MF were significantly enriched in cyclin binding, ATPase activity, and growth factor activity. (Figure 2c). Result of KEGG analysis showed that top pathways were cell cycle, p53 signaling pathway, TNF signaling pathway, cytosolic DNA‐sensing pathway, Oocyte meiosis (Figure 2d).

FIGURE 2

The enrichment analysis of DEGs by DAVID and Metascape. Detailed information relating to changes in the (a) CC, (b) BP, (c) MF, and (d) KEGG analysis for hub genes. (e) Heatmap of enriched terms across input differently expressed gene lists, colored by p‐values, via the Metascape. (f) Network of enriched terms colored by cluster identity, where nodes that share the same cluster identity are typically close to each other. (g) Network of enriched terms colored by p‐value, where terms containing more genes tend to have a more significant p‐value

The analysis of metascape

Furthermore, the functional enrichment analysis with Metascape manifested that the DEGs between TS and un‐TS samples were significantly enriched in mitotic cell cycle process, cytokinetic process, cell cycle, formation of the cornified envelope, monocyte chemotaxis, meiotic nuclear division, cytosolic DNA‐sensing pathway, SUMOylation of DNA replication proteins, inflammatory response, negative regulation of proteolysis, chromosome condensation, regulation of growth, ovulation, protein localization to cytoskeleton (p < .05, Figure 2e–g).

The identification of significant module

Four significant modules were also identified by the MCODE (Figure 3a–d).

FIGURE 3

The four significant modules identified from the PPI network

The identification and analysis of hub genes

The 10 hub genes were identified by “cytoHubba” (Figure 4a). Heatmap showed that hub genes could differentiate the TS and un‐TS samples. When compared with un‐TS samples, the hub genes were upregulated in the TS samples (Figure 4b). The Pearson correlation analysis manifested that there exist positive correlations among all the hub genes (Figure 5a).

FIGURE 4

The hub genes network via cytoHubba_Degree (a), and the heatmap presenting the expression of these hub genes (b)

FIGURE 5

(a) The Pearson correlation analysis manifested that there exist positive correlations among these hub genes via cytoHubba_Degree. (b) The hub genes network via cytoHubba_MCC. (c) The VENN diagram between “cytoHubba_Degree” and “cytoHubba_MCC.” (d) Relationship with trachea related to MKI67 based on the CTD database. (e) Relationship with trachea related to CCNB1 based on the CTD database. (f) Relationship with trachea related to CCNB2 based on the CTD database. Note: MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10)

The most significant genes

After analysis of MCC, 10genes were selected from PPI network (Figure 5b). The VENN diagram showed that there were three most significant genes (Mki67, Ccnb1, Ccnb2) between “cytoHubba_Degree” and “cytoHubba_MCC” (Figure 5c).

The CTD analysis of most significant genes

Identification of significant genes associated with tracheal diseases, cartilage diseases, tracheal neoplasms, fibroadenoma, and connective tissue diseases was performed on CTD database which is shown in Figure 5d–f. Specifically, the MKI67 is shown in Figure 5d, the CCNB1 in Figure 5e, and the CCNB2 in Figure 5f.

Patient characteristics

The demographic data and the expression status of MKI67, CCNB1, and CCNB2 are summarized in the Table 2. The mean patient age was 54 years (range, 25–85 years).

TABLE 2

Clinical variables and expression status of MKI67, CCNB1, and CCNB2

Characteristics	Diagnosis		p‐value
	Tracheal stenosis	Control
Sex _ Male	10 (40.0%)	10 (40.0%)	.227
Sex _ Female	4 (16.0%)	1 (4.0%)
Age (years old)	47.36 ± 18.44	59.64 ± 14.65	.076
Height (cm)	169.82 ± 3.68	165.64 ± 5.84	.050
Weight (kg)	74.96 ± 8.43	73.00 ± 9.88	.606
MKI67	2.42 ± 0.59	0.61 ± 0.39	.000
CCNB1	4.14 ± 0.92	1.44 ± 0.36	.000
CCNB2	3.64 ± 0.55	0.78 ± 0.23	.000

MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10).

RT‐qPCR analysis validation of the most significant genes (MKI67, CCNB1, and CCNB2)

As presented in the result, the expression of MKI67 (p < .05, Figure 6a), CCNB1 (p < .05, Figure 6b), and CCNB2 (p < .05, Figure 6c) were significantly upregulated in TS samples. It should be noted that the expression situation of MKI67, CCNB1, and CCNB2 were consistent in above bioinformatic results. The degree of TS in the TS group was higher (Figure 6d).

FIGURE 6

RT‐qPCR analysis validation of the most significant genes, (a) MKI67, (b) CCNB1, and (c) CCNB2. (d) Comparison of degree of TS between TS and control groups. Note: MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10)

Strong positive associations between the degree of TS, MKI67, CCNB1, and CCNB2

The degree of TS was positively associated with the MKI67 (R = .906, p < .001; Figure 7a). And the degree of TS was positively related with CCNB1 (R = .953, p < .001; Figure 7b). The degree of TS was positively associated with the CCNB2 (R = .958, p < .001; Figure 7c). And MKI67 was positively associated with CCNB1 (R = .912, p < .001; Figure 7d). The relative expression of MKI67 was positively related with CCNB2 (R = .862, p < .001; Figure 7e). CCNB1 was positively associated with CCNB2 (R = .936, p < .001; Figure 7f). The heatmap showed the strong associations between the MKI67, CCNB1, and CCNB2 through the Spearman correlation analysis (Figure 8).

FIGURE 7

Strong positive associations between the degree of TS, MKI67, CCNB1, and CCNB2 based on the Pearson correlation analysis. Note: MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10)

FIGURE 8

The heatmap presenting the correlation among the MKI67, CCNB1, and CCNB2 based on the Spearman correlation analysis. Note: MKI67 (Reference Sequence: NG_047061.1), CCNB1 (Reference Sequence: NC_000005.10), CCNB2 (Reference Sequence: NC_000015.10)

DISCUSSION

Tracheal injury can cause intratracheal scar hyperplasia which in turn causes TS (Farzanegan et al., 2017). With the increasing use of mechanical ventilation and ventilator, the incidence of TS is increasing (Kim et al., 2017). What is more, Patients with TS may experience symptoms such as dyspnea, wheezing, and other symptoms and their life and work are seriously affected (Farzanegan et al., 2017). A retrospective research found that 54.7% of the patients with TS had iatrogenic causes (Gelbard et al., 2015). Meanwhile, aging, obesity, and diabetes are also risk factors for TS (Li et al., 2018). However, the molecular mechanisms of TS have not been elucidated. And the diagnosis of TS is usually made after discharge, affecting the therapeutic effect (Shadmehr et al., 2017). Therefore, it is important to find the pathogenesis of TS for the early diagnosis and targeted therapy. Bioinformatics methods can be used to explore DEGs between TS and normal individuals. Through further analysis, multiple hub genes were found, of which MKI67, CCNB1, and CCNB2 are worth paying attention.

MKI67, the marker of proliferation Ki‐67, is mainly involved in regulating protein binding, DNA binding, regulation in mitotic nuclear division, cell proliferation, and regulation of chromosome segregation. What is more, MKI67 might be involved in the occurrence of many diseases. Tokarz found that overexpression of MKI67 was accompanied by inflammatory stress and cell death, suggesting that MKI67 may be involved in the process of cell cycle and proliferation (Tokarz, Piastowska‐Ciesielska, Kaarniranta, & Blasiak, 2016). Hou found that knocking out MKI67 can inhibit the growth of cancer cells, suggesting that MKI67 may take part in cell division, and then affect the cell cycle (Hou et al., 2011). Similarly, through gene sequencing and association analysis, Yang suggested that MKI67 and TP53(OMIM: 191170) may serve as molecular markers for cell cycle regulation (Yang et al., 2017). Xiong retrospectively analyzed multiple studies of 7,078 patients and found that high expression of MKI67 reduced the overall survival rate and tumor‐free survival rate. Further functional analysis revealed that MKI67 might take part in the occurrence of apoptosis by regulating the p53 signaling pathway (Xiong, Zeng, Jiang, Luo, & Chen, 2019). Furthermore, Zeng demonstrated that SNRPA1 (OMIM: 603521) can regulate PIK3R1 (OMIM:171833), VEGFC (OMIM:601528), and MKI67 molecules in colorectal cancer through gene overexpression and gene knockout experiments, and then be involved in tumorigenesis and development, suggesting that MKI67 may serve as a molecular target for cell cycle and apoptosis (Zeng et al., 2019). Consistent with the above studies, we found that MKI67 is highly expressed in TS tissues. We speculated that MKI67 may promote tracheal endothelial cell proliferation and fibrosis by affecting the cell cycle, inducing scar tissue hyperplasia and then TS. MKI67 might serve as a target for specific therapy of TS, and the relevant mechanisms are worth further exploration.

CCNB1, cyclin B1, is mainly involved in the cell cycle. Abnormal expression of CCNB1 might lead to various diseases. Liu found that the polymorphism of CCNB1 was significantly related with the efficacy of platinum‐based chemotherapy. The polymorphism of CCNB1 may be involved in the gastrointestinal toxicity of platinum‐based chemotherapy. Further analysis suggests that CCNB1 might regulate the cell cycle and G2/M phase, providing new ideas for individual treatment of patients with non–small‐cell lung cancer (Liu et al., 2017). What is more, Zhang found that CCNB1 was highly expressed in pancreatic cancer tissues by quantitative detection. CCNB1 might affect pancreatic cancer cell cycle, proliferation and apoptosis through TP53 pathway (Zhang, Feng, et al., 2018; Zhang, Zhang, et al., 2018). In addition, Gu found that inhibition of CCNB1 expression could significantly inhibit the proliferation, migration, and invasion of cells (Gu, Liu, Li, & He, 2019). Li found that CCNB1 could affect the proliferation and apoptosis of pituitary adenoma cells and activate epithelial‐to‐mesenchymal transition. Knockout of the CCNB1 gene could inhibit cell proliferation (Li et al., 2019), suggesting that CCNB1 could regulate cell cycle and affect cell proliferation. Furthermore, CCNB1 is involved in regulating autophagy (Lu et al., 2019), which in turn regulates the degree of tissue fibrosis (Qin et al., 2017). Apoptosis of tracheal epithelial cells can promote the proliferation of fibroblasts, thereby aggravating TS (Wang et al., 2016). Consistent with the above studies, we found that CCNB1 is highly expressed in TS tissues. We speculate that CCNB1 may be involved in the occurrence and development of TS by regulating cell cycle, autophagy, and apoptosis. Overexpression of CCNB1 can accelerate the cell cycle and promote endothelial proliferation, fibrosis, and hornification, then inducing TS. CCNB1 may serve as a target for early diagnosis and specific therapy of tracheal stricture, and the related mechanisms need to be further studied.

CCNB2, cyclin B2, is mainly involved in the regulation of growth, T‐cell homeostasis, and cell cycle. Ian found significant overexpression of CCNB2 in non–small‐cell lung cancer tissues through bioinformatics analysis and quantitative validation, and the high expression of CCNB2 protein can affect the overall survival rate of non–small‐cell lung cancer patients, suggesting that CCN B2 may be involved in the development of non–small cell lung cancer by regulating G2/M transition of cell cycle (Qian et al., 2015). What is more, Shi found that Islet‐1 can regulate the expression of CCNB2, and then promote cell cycle progression of gastric cancer (GC) cells. In this study, the proportion of cells in the G1 phase is reduced, and the proportion of cells in the G2/M and S phases is increased, thereby promoting the proliferation of GC cells and providing new evidence for the study of the mechanisms of GC (Shi et al., 2016). In addition, Gao found that KPNA2 (OMIM: 600685) can upregulate the expression of CCNB2, and then induce the proliferation of hepatoma cells, thereby promoting the development of hepatoma, suggesting that related molecules can serve as therapeutic targets (Gao, Wang, Yang, Yang, & Zhuang, 2018). Furthermore, Wang found multiple genes differentially expressed in scar tissues and normal tissues through bioinformatics analysis. And through further analysis, Wang found that CCNB2 may be involved in the development of scar by regulating the TP53 signaling pathway, providing ideas for the study of the mechanisms of scar (Zeng et al., 2019). Consistent with the above studies, we found that CCNB2 was highly expressed in TS. We speculate that CCNB2 may be involved in the occurrence and development of TS by regulating cell cycle, inflammatory factors and inducing airway epithelial cell proliferation, fibrosis, and scar formation. CCNB2 may serve as a target for early diagnosis and specific therapy of tracheal stricture, and the related molecular mechanisms need to be further clarified.

However, there were some shortcomings. The small sample size in the dataset still needs to be further expanded to obtain more accurate results to better understand the molecular mechanisms of TS.

In summary, there are many DEGs between TS and normal tracheal tissues, especially MKI67, CCNB1, and CCNB2. These molecules might be molecular targets for early diagnosis and specific therapy of TS.

CONFLICT OF INTEREST

The authors declare they have no conflict of interest.

AUTHOR CONTRIBUTION

Conceptualization, Methodology: Lining Huang, Wen‐ping Song. Data curation, Formal analysis, Investigation: Yong Qiu, Shu‐xian Ma. Project administration, Resources, Software: Wen‐hao Wu, Pei Zhang. Validation, Visualization: Ling‐bing Meng, Wei Yang. Writing‐original draft, Writing‐review & editing: Xu‐ze Li, Zi‐chen Wang. All authors read and approved the final manuscript.